XAS and XRD study of annealed 238Pu- and 239Pu-substituted zircons (Zr0.92Pu0.08SiO4)

Journal of Nuclear Materials 278 (2000) 212±224

XAS and XRD study of annealed 238Pu- and zircons (Zr0:92Pu0:08SiO4)

www.elsevier.nl/locate/jnucmat

239

Pu-substituted

B.D. Begg a, N.J. Hess b, W.J. Weber b,*, S.D. Conradson c, M.J. Schweiger b, R.C. Ewing d a

Australian Nuclear Science and Technology Organisation, PMB 1, Menai, NSW 2234, Australia Paci®c Northwest National Laboratory, PO Box 999, MSIN K2-44, Richland, WA 99352, USA c Los Alamos National Laboratory, Los Alamos, NM 87545, USA Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, MI 48109, USA b

d

Received 1 June 1999; accepted 29 September 1999

Abstract An annealing study has been completed on two compositionally identical Pu-substituted zircons (Zr0:92 Pu0:08 SiO4 ) prepared 18 yr ago with di€erent 238 Pu/239 Pu isotopic ratios. The activity and accumulated dose for each material varies greatly due to the very di€erent half-lives of 238 Pu (87.7 yr) and 239 Pu (24 100 yr). The predominantly 238 Pu-substituted zircon is in a highly amorphous state after accumulating a dose of 2:8  1019 a-decays/g, which is a factor of two higher than the dose previously determined necessary to amorphize this material. The 239 Pu-substituted zircon has remained highly crystalline after an accumulated dose of 1:2  1017 a-decays/g. The Pu in both samples is trivalent due to sintering under reducing conditions. The short-range and long-range structures of each sample have been characterized by detailed X-ray absorption spectroscopy and X-ray di€raction (XRD) methods. The oxidation of Pu3‡ to Pu4‡ in the crystalline 239 Pu-substituted zircon during annealing in air resulted in a decrease in lattice distortion due to the decrease in ionic radius of Pu3‡ to Pu4‡ on the Zr4‡ site. A signi®cant amount of PuO2 exsolved from the zircon structure after annealing at 1200°C in air. Detailed characterization of the amorphous 238 Pu-substituted zircon demonstrated that while devoid of long-range order it still retained a distorted zircon structure and composition over a length scale of 0.5 nm. The recrystallization of the amorphous 238 Pu-substituted zircon could be achieved directly at temperatures as low as 1200°C when annealed for 12 h in air; however, annealing at 1000°C resulted in decomposition into constituent oxides. Ó 2000 Elsevier Science B.V. All rights reserved. PACS: 61.10.Ht; 61.10.Nz; 61.43.Er; 61.80.-x; 61.82.Ms

1. Introduction One of the principal factors that will determine the long-term durability of nuclear waste forms proposed for the immobilization of high-level radioactive waste or actinide-rich waste is the structural response to self-radiation damage as a function of time and temperature.

*

Corresponding author. Tel.: +1-509 375 2299; fax: +1-509 375 2186. E-mail address: [email protected] (W.J. Weber).

This is especially true for waste forms designed to incorporate actinide elements, such as plutonium, because they will be subject to considerable self-radiation damage from a-decay events [1,2]. With this in mind, an accelerated study of a-decay damage was initiated 18 yr ago to investigate self-radiation e€ects from a-decay in 238 Pu-substituted zircon, ZrSiO4 , and compare the results with data on natural zircons. The detailed results of damage accumulation and recovery from the initial phase of this study have been reported previously [3±6]. Although the initial study focused primarily on 238 Pusubstituted zircon, two types of synthetic zircons were

0022-3115/00/$ - see front matter Ó 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 3 1 1 5 ( 9 9 ) 0 0 2 5 6 - 1

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

actually prepared, one containing a mixture of 8.85 wt% 238 Pu and 1.15 wt% 239 Pu and the other containing 10 wt% 239 Pu. In both instances, the Pu was substituted directly for Zr, giving a Zr0:92 Pu0:08 SiO4 stoichiometry. Initial characterization by X-ray di€raction (XRD), found both samples to be single-phase zircons. The ®rst zircon containing 8.85 wt% 238 Pu, with its 87.7 yr halflife, provided a means of accelerating the a-decay rate (i.e., damage rate) by a factor of 250 when compared to the sample containing 10 wt% 239 Pu. This 238 Pu-substituted zircon became amorphous to XRD analysis after a dose of 0:7  1019 a-decays/g [3,5]. Details of the sample preparation procedures and the initial characterization have been reported previously [3±6]. Selfheating in the 238 Pu-substituted zircon specimens during storage is minimal, and specimen temperatures have been estimated to be less than 50°C. These zircon samples have been more recently characterized by X-ray absorption spectroscopy (XAS) after accumulated doses of 2:8  1019 and 1:2  1017 a-decays/ g for the 238 Pu- and 239 Pu-substituted zircons, respectively [7,8]. The 238 Pu-substituted zircon in these studies achieved a dose that is a factor of two higher than that determined necessary to fully amorphize the same samples [3±6]. The XAS results [7,8] con®rmed that self-radiation damage had transformed the 238 Pu-substituted zircon into a fully amorphous state lacking long-range order. Surprisingly, the Pu LIII -edge XANES indicated that the Pu in both zircon samples was completely trivalent. This was an unexpected result of originally preparing the Pu-substituted zircons under a reducing atmosphere (oxygen-purged ¯owing argon). The trivalent state of Pu in zircon poses a number of questions, notably how is Pu3‡ incorporated into the zircon structure, due to the apparent signi®cant ionic radii mismatch between Pu3‡ , 0.100 nm (sixfold coordination [9]) and Zr4‡ , 0.084 nm (eightfold coordination [9]), and consequently how has charge neutrality been maintained? Although it is preferable to compare the ionic radii of Pu3‡ and Zr4‡ on equivalently coordinated sites, the ionic radius of Pu3‡ in eightfold coordination corresponding to the Zr4‡ site in zircon is unknown. Recent computer simulations, however, using energy minimization techniques [10] indicate that the lowest energy con®guration occurs for a defect cluster composed of two near-neighbor Pu3‡ substitutions on Zr4‡ sites and a neighboring charge-compensating oxygen vacancy. In order to further understand the behavior of Pu in zircon and recrystallization processes, an annealing study was undertaken with the twofold aim of oxidizing (in air) the Pu3‡ to Pu4‡ in both the amorphous (238 Pu-substituted) and crystalline (239 Pu-substituted) zircons, as well as to examine the nature of the recrystallization processes in the amorphous state of zircon. The original goal was to oxidize the Pu3‡ in the amorphous zircon at a suciently low temperature to avoid any recrystallization.

213

There have been several previous isochronal annealing studies on zircons that are relevant to the present study. One of these studies [3,5,6] was carried out on the same 238 Pu-substituted zircon as in the current study but at a lower dose (1:35  1019 a-decays/g) where this zircon is ®rst observed to be fully amorphous. Other studies have been carried out on natural metamict zircons [11±17]. Since zircon undergoes a large (18%) volume expansion upon amorphization [3±6], the isochronal recovery processes in 238 Pu-substituted zircon [3,5,6] and in natural metamict zircon [11] were monitored via density measurements, as shown in Fig. 1. In each of these studies, a single specimen was sequentially annealed to successively higher temperatures. After each heat treatment, the sampleÕs density was determined, and the structure was monitored by XRD methods or in some instances by optical spectroscopy. The ®rst stage of the two-stage density recovery process is consistent with the initial crystallization of zirconÕs component oxides and would account for the observed 55±60% recovery of the density in both studies. The second stage of the density recovery process, which occurs above 1200°C, has been only observed in the 238 Pu-substituted zircon [3,5,6] and represents the recrystallization of the zircon structure, which was complete at 1450°C. Although the second stage of the density recovery process in the natural metamict zircon was not observed due to a restricted maximum temperature [11], Vance and Anderson [18] have reported the recrystallization of the zircon structure in another natural metamict zircon at 1300°C. The isochronal density recovery curves from the two studies (Fig. 1) are in good agreement with each other and consistent with the recovery of the integrated neutron di€raction intensity in natural metamict zircon [14]. The lower onset temperature for recovery that is

Fig. 1. Summary of isochronal density recovery studies of amorphous zircons by Weber [3,5,6] and Vance and Anderson [11].

214

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

observed for the 238 Pu-substituted zircon occurs because a 12 h isochronal period was used in the annealing of the 238 Pu-substituted zircon, while only a 1 h isochronal period was used for the metamict zircons. In the natural metamict zircon, Vance and Anderson [11] noted the development of di€use di€raction maxima associated cubic zirconia after annealing at 850°C. By 1100°C, the di€use pattern converted to the sharp di€raction pattern of tetragonal zirconia. At temperatures greater than 1150°C, the intensity of the tetragonal zirconia peaks decreased, although no evidence for zircon was observed in the powder XRD pattern at the maximum temperature of 1350°C. However, di€use streaks in a Laue diffraction pattern, which had the symmetry of zircon, were observed after annealing at 1350°C. In the 238 Pusubstituted zircon [3,5,6], the appearance of cubic zirconia was ®rst noted after annealing at 1050°C. In light of these previous studies, the amorphous 238 Pu-substituted zircon was annealed at 1000°C for 12 h in air. This temperature represented a compromise between being suciently high enough to ensure complete oxidation of the Pu3‡ and low enough to avoid recrystallization. A second 238 Pu-substituted zircon specimen was annealed at 1200°C for 12 h in air in order to study the recrystallization of the amorphous zircon in the transition region of the two-stage recovery process. In addition, the crystalline 239 Pu-substituted zircon was annealed at 1200°C for 12 h in air in order to oxidize the Pu3‡ . In these annealing studies, the heating rate was on the order of 600°C/h.

at 77 K. The transmission spectra were measured with standard ionization chamber detectors, while the ¯uorescence was recorded with a 13-element Ge detector. Energy calibration was performed by assigning the ®rst in¯ection point in the absorption edge of a Zr foil to 17,999.35 eV. Data analysis was performed in a standard manner that has been described previously [7,8]. The XRD spectra were collected in transmission out to a maximum of 2 nmÿ1 in reciprocal space using monochromatic X-rays of energy between 20±30 keV, at 77 K. A solid-state detector coupled to a digital ampli®er, which provided an energy spectrum at every point in the pattern that is analyzed by the data acquisition software, was used to measure the di€racted beam intensity, as well as the sampleÕs ¯uorescence that enabled volume corrections to be made. The intensity of the incident beam was also monitored with an ionization chamber to provide for normalization. Pair distribution function analysis of the di€raction pattern was carried out using PDF 1.00 software [19,20]. 3. Results 3.1. 239 Pu-substituted zircon annealed at 1200°C for 12 h in air Pu LIII -edge XANES revealed that annealing the Pu-substituted zircon at 1200°C for 12 h in air was sucient to fully oxidize the Pu3‡ to Pu4‡ . XRD indicated, however, that as a result of this heat treatment the sample now contained a considerable amount of PuO2 in addition to zircon. This was quite unexpected given that the unannealed 239 Pu-substituted zircon was single phase, and the Pu LIII -edge EXAFS had indicated that the Pu was uniformly distributed within the zircon lattice, which excluded the possibility of any amorphous Pu-bearing second phases in the unannealed zircon [7,8]. Therefore, the PuO2 observed in the annealed 239 Pusubstituted zircon appears to have been exsolved from the zircon lattice as a result of this 1200°C heat treatment. The lattice parameters obtained from the zircon before and after annealing are compared in Table 1, along with the published reference values for zircon [21]. As expected, the unannealed 239 Pu-substituted zircon, in which the Pu was completely trivalent and contained 239

2. Experimental The original preparation of the 238 Pu- and 239 Pusubstituted zircon samples used in this study has been described previously [3,5,6]. The high-temperature anneals were carried out on small (1 g) solid specimens that were subsequently ground into powder, mixed with colloidion-amyl acetate solution, and loaded as a slurry into aluminium holders with Kapton windows. EXAFS and XRD measurements were performed at the Stanford Synchrotron Radiation Laboratory. EXAFS spectra were collected at the Pu LIII - and Zr K-edges in transmission and ¯uorescence modes simultaneously out to a photoelectron wavevector of 1.3 nmÿ1

Table 1 A comparison of the unit cell volumes obtained from the 239 Pu-substituted zircon before and after annealing at 1200°C for 12 h in air with those for a reference zircon [21] 239 3

Cell volume (nm ) D rel. ZrSiO4 Pu valence Ionic size (nm) % Pu in zircon

Pu-zircon

0.2678 +0.0070 +3 0.100 (Pu) 100%

239

Pu-zircon 1200°C/12 h

0.2623 +0.0015 +4 0.086 (Pu) 50%

ZrSiO4 [21] 0.2608 0 ± 0.072 (Zr) 0%

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

in the zircon structure, had the largest unit-cell volume. The oxidation of the Pu3‡ to Pu4‡ combined with the exsolution of Pu from the annealed 239 Pu-substituted zircon lattice considerably reduced its unit-cell volume. A ®rst-order estimate of the amount of Pu exsolved from the annealed 239 Pu-substituted zircon may be made, based on the measured unit-cell volumes and the differences in ionic radii between Pu3‡ , Pu4‡ and Zr4‡ , as outlined in Table 1. Since the di€erence in ionic radii between Pu3‡ and Zr4‡ is twice that between Pu4‡ and Zr4‡ , it would be expected that, relative to the reference zircon, the increase in unit-cell volume for a given amount of Pu3‡ substituted on the Zr4‡ site should be approximately twice that for Pu4‡ . Given that the measured di€erence in the unit-cell volume of the annealed 239 Pu-substituted zircon, relative to the reference zircon, is only 43% of that expected, the annealed 239 Pu-substituted zircon is estimated to contain 43% of the Pu. Recent Mean Field calculations [10], however, have predicted that the increase in lattice volume associated with the substitution of an equivalent amount of Pu3‡ instead of Pu4‡ on the Zr4‡ site would be a factor of 1.8 rather than 2, as used here. Based on this value, the amount of Pu retained in the zircon is estimated to be 48%. A number of other factors also in¯uence the unit-cell volumes that have not been taken into account in this ®rst order approximation. These include the presence of oxygen vacancies, which would be the most likely means of charge compensating for the presence of Pu3‡ on the Zr4‡ site in the unannealed 239 Pu-substituted zircon, and radiation damage induced expansion of the unit cell. The later e€ect has, however, been shown to be almost insigni®cant for the low dose, 1:2  1017 a-decays/g, experienced by the 239 Pu-substituted zircon [5]. In light of the uncertainty in both the measurements and the calculations, it would be safe to conclude that the annealed 239 Pu-substituted zircon only contains approximately half of the original 0.08 formula units of Pu. A comparison of the Fourier transforms of the Pu LIII -edge EXAFS collected before and after annealing is shown in Fig. 2. The presence of a signi®cant amount of PuO2 is clearly evident after the oxidizing anneal that has resulted in the large increase in the intensity of many of the transform features. In addition to the contribution from PuO2 , there is also an increase in the intensity of the resolvable Pu±Si shell from the zircon. An increase in the modulus of the Pu±Si feature in the transform indicates that the level of distortion associated with the presence of Pu in the zircon is reduced after the oxidizing anneal. This is expected given the reduction in the ionic size mismatch associated with the oxidation of the Pu3‡ to Pu4‡ on the Zr4‡ site and the ®lling of the closely associated oxygen vacancy site. This decrease in the level of distortion may also be partly associated with the annealing of any irradiation-induced defects that may have been created at this low dose level.

215

Fig. 2. Fourier transforms of the Pu LIII -edge EXAFS taken from the 239 Pu-substituted zircon before and after annealing in air at 1200°C for 12 h. Note that peaks in the Fourier transform have not been corrected for phase shift and as a result appear 0.05±0.10 nm shorter than ®tted values.

A comparison of the Pu LIII -edge EXAFS ®tting results obtained from the 239 Pu-substituted zircon before and after annealing, is shown in Table 2. The Pu environment in the unannealed 239 Pu-substituted zircon is well described by the PuSiO4 standard [22], which would suggest that the Pu3‡ present in the unannealed 239 Pusubstituted zircon has substituted on the Zr4‡ site as planned, despite the large ionic size mismatch. Charge compensation has most likely occurred through the presence of an appropriate number of oxygen vacancies, giving the zircon [Zr0:92 Pu0:08 ]Si[O3:96 h0:04 ] stoichiometry. This is consistent with recent theoretical defect calculations [10] that indicated that the energy cost associated with placing Pu3‡ in an interstitial position is ®ve times that for placing it on the Zr4‡ site. In addition, the calculations predict that the defect energy can be reduced by a further factor of four, if the Pu3‡ forms clusters consisting of two Pu3‡ on near-neighbor Zr4‡ sites and an oxygen vacancy for charge compensation. No evidence for Pu clustering (i.e., as PuO2 ) was observed in the Pu LIII -edge EXAFS from the unannealed 239 Pu-substituted zircon. As indicated by the Fourier transform, PuO2 dominated the ®t to the Pu LIII -edge EXAFS from the annealed 239 Pu-substituted zircon. After the anneal, the intensity of the resolvable Pu±Si shell both increased and appeared at a signi®cantly shorter distance, which would be expected given the reduction in ionic size associated with the oxidation of Pu3‡ to Pu4‡ . A comparison of the local Zr environments from the 239 Pu-substituted zircon before and after annealing at 1200°C in air is shown in Fig. 3. The Fourier transforms of the Zr K-edge EXAFS indicate good structural

216

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

Table 2 A comparison of the Pu LIII -edge EXAFS ®tting results for the 239 Pu-substituted zircon before and after being annealed at 1200°C for 12 h in air. Also shown are reference data from PuSiO4 [22] and PuO2 [29] Shells

Pu±O

Pu±O

Pu±Si

Pu±Si/Zr

Pu±O

Distance (nm) Number D Sigma

0.221 (2) 2.9 (8) 0.00

0.239 (2) 5 (1) 0.00

0.316 (2) 1.0 (3) 0.07 (2)

0.364 (2) 7 (2) 0.14 (2)

0.415 (2) 3 (1) 0.00

PuSiO4

Distance (nm) Number

0.229 4

0.249 4

0.311 2

0.378 4/4

0.421 8

Shells

Pu±O

Pu±Si

Pu±Pu

Pu±O

Pu±Pu

239

Distance (nm) Number D Sigma Distance (nm) Number

0.230 (2) 7 (2) 0.08 0.234 8

0.300 1.6 (5) 0.07

0.387 (1) 6 (1) 0.02 0.382 12

0.438 (2) 23 (6) 0.03 0.447 24

0.532 (2) 3 (1) 0.00 0.539 6

239

Pu-zircon

Pu-zircon 1200°C/12 h

PuO2

reduction of the ®rst Zr±Si correlation in the unannealed 239 Pu-substituted zircon may indicate disruption of some edge-sharing relationships, formation of some cornersharing con®gurations by the rotation of SiO4 tetrahedra away from the ZrO8 polyhedra, and creation of some oxygen vacancies for the charge compensation of the Pu3‡ . The extent of atom pair correlations was determined by taking the Fourier transform of the corrected diffraction pattern to give a pair distribution function (PDF). The PDF shows the probability of ®nding two atoms a distance r apart weighted by the product of their atomic numbers. Comparisons of the PDF patterns taken before and after annealing are shown in Fig. 4. While the common features of zircon are clearly evident, the PDF features from the annealed 239 Pu-substituted Fig. 3. Fourier transforms of the Zr K-edge EXAFS taken from the 239 Pu-substituted zircon before and after annealing in air at 1200°C for 12 h. Note that peaks in the Fourier transform have not been corrected for phase shift and as a result appear 0.05±0.10 nm shorter than ®tted values.

agreement out to 0.65 nm, beyond which deviations exist. Having said that, a signi®cant di€erence is present in the region of 0.25 nm, where the ®rst resolvable Zr± Si distances reside. The e€ect of the oxidation of the Pu3‡ to Pu4‡ , and the accompanying reduction in lattice distortion, is therefore clearly evident on the neighboring Zr sublattice through changes in the ®rst Zr±Si correlation, which appears very sensitive to structural distortion in the zircon. Zircon is composed of chains of alternating edge-sharing ZrO8 polyhedra and SiO4 tetrahedra that are cross-linked through corner-sharing SiO4 tetrahedra. The ®rst Zr±Si correlation at 0.299 nm arises from edge-sharing ZrO8 polyhedra and SiO4 tetrahedra; whereas, the Zr±Si correlation at 0.366 nm results from corner sharing polyhedra and tetrahedra. The

Fig. 4. Pair distribution function obtained from the 239 Pusubstituted zircon before and after annealing in air at 1200°C for 12 h. The positions of the dominant Pu±Pu pairs from PuO2 are indicated by arrows.

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

zircon are signi®cantly sharper and more intense than those from the unannealed 239 Pu-substituted zircon, re¯ecting the signi®cant reduction in lattice distortion associated with the oxidation of the Pu3‡ to Pu4‡ and recovery of dilute damage from the relatively small adecay dose. The intensity variations seen in the PDF data are consistent with those observed in the Fourier transforms from the Pu and Zr environments described above and re¯ect the complementary nature of these independent measurements. The presence of PuO2 in the annealed 239 Pu-substituted zircon was also evident, as indicated by the location of the dominant Pu±Pu pairs from PuO2 in Fig. 4. 3.2.

238

Pu-substituted zircon

Prior to discussing the annealing studies on the amorphous 238 Pu-substituted zircon, it is important to review the earlier XAS ®nding [7,8] in light of more recent XRD results. Fig. 5 shows a normalized coherent di€raction pattern obtained from the 238 Pu-substituted zircon using high-brightness synchrotron radiation. The very broad di€use nature of the di€raction pattern is consistent with the lack of long-range order in the zircon. Superimposed on the amorphous background, however, is a very small amount of PuO2 , which is estimated to be less than 1% of the total Pu present in the sample. This was most likely present in the starting material but, due to its low concentration, was not detectable above the initial crystalline zircon background with the laboratory-source XRD unit used in the initial study [3]. The presence of such a small proportion of PuO2 is also not irreconcilable with the earlier Pu LIII edge XANES ®ndings that indicated that all the Pu was trivalent due to the limited sensitivity of the technique.

Fig. 5. Corrected XRD pattern from the amorphous (unannealed) 238 Pu-substituted zircon, showing a small amount of PuO2 .

217

PDF analysis of this di€raction pattern was undertaken to look for evidence of short-range order in this otherwise amorphous material, and the results are shown in Fig. 6. Also shown is a ®t to the data based on crystalline zircon to which a broadening component has been added to simulate lattice distortion. Despite its amorphous state, there appears to be excellent agreement between the data and the distorted zircon ®t out to a distance of 0.5 nm, after which the long-range zircon structure appears to have been lost. The two minor peaks in the PDF at 0.66 nm and 0.85 nm coincide with the Pu±Pu distances in PuO2 , indicated in Fig. 6. A detailed breakdown of the cation pairings that constitute the ®t out to 0.5 nm, also shown in Fig. 6, indicates that despite the loss of long-range crystallinity a distorted zircon structure remains over a length scale of about 0.5 nm. This is consistent with the retention of the ®rst Zr±Zr and Si±Si pairs. The ®rst Zr±Si pairs at 0.3 nm appear to have been lost, which suggests a complete loss of the edge-sharing relationships and the rotation of the SiO4 tetrahedra away from the ZrO8 polyhedra in the amorphous state, as suggested by Farges and Calas for metamict natural zircon [23] and similar to that proposed for cation polyhedra in natural metamict zirconolite [24]. The ®t also suggests that the main Zr±Zr correlation has contracted to a slightly shorter distance, which is consistent with previous EXAFS results on these samples [7] and on natural metamict zircons [15,23]. The local Pu environment in this sample was characterized with Pu LIII -edge EXAFS, and the ®tting

Fig. 6. PDF plot from amorphous (unannealed) 238 Pu-substituted zircon, including a ®t to the data based on crystalline zircon to which a broadening term has been applied to account for lattice distortion. The detailed ®t shows the contributions from the various cation±cation interactions out to 0.5 nm. Two minor peaks that coincide with the Pu±Pu distances in PuO2 are indicated.

218

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

results are shown in Table 3. A comparison of the ®tting results with the PuSiO4 standard [22] indicates that all the cation±cation pairings present in the Pu end-member zircon are present in this nominally amorphous 238 Pusubstituted zircon out to a distance of 0.5 nm, in agreement with the PDF result. No evidence for either amorphous Pu-bearing second phases or Pu clustering within the zircon was indicated by the Pu LIII -edge EXAFS data. Notable distortion was, however, present through the reduction in the number of the atom pairs, the presence of only one Pu±O distance at 0.23 nm, compared to the two in PuSiO4 , and the contraction of the Pu±Zr distance to 0.36 nm as indicated from the PDF. However, despite the loss of long-range crystallinity through the signi®cant accumulation of a-decay damage, the local Pu environment out to 0.5 nm remains well described by PuSiO4 and consequently is remarkably similar to the Pu environment in the crystalline 239 Pu-substituted zircon described in Table 2. Further, the retention of a distorted zircon structure around the Pu indicates that the zircon stoichiometry has been retained on the atomic level in the amorphous 238 Pu-substituted zircon, similar to results reported for natural metamict zircons [12,23]. The ®tting results from the Zr K-edge EXAFS from the unannealed 238 Pu-substituted zircon are shown in Table 4, along with reference distances for crystalline Table 3 Pu LIII -edge EXAFS ®tting results from the unannealed 238

Pu-zircon

PuSiO4

Pu-zircon

ZrSiO4

ZrO2 (M)

Pu-substituted zircon, along with reference data from PuSiO4 [22]

Shells

Pu±O

Pu±O

Pu±Si

Pu±Zr

Pu±Si

Pu±Si

Distance (nm) Number DSigma

0.232 (4) 5 (1) 0.17 (2)

± ± ±

0.316 (3) 1.7 (5) 0.11 (2)

0.360 1.3 (4) 0.10 (1)

0.383 0.9 (3) 0.11 (2)

0.443 3 (1) 0.16 (2)

Shells

Pu±O

Pu±O

Pu±Si

Pu±Si/Zr

Pu±O

Pu±Si

Distance (nm)

0.229

0.249

0.311

0.378

0.421

0.488

Table 4 Zr K-edge EXAFS ®tting results from the [25] 238

238

zircon [21] and monoclinic ZrO2 [25]. A number of structural deviations exist, including the presence of a single Zr±O distance at 0.22 nm, which is the average of the two Zr±O distances in crystalline zircon. The presence of two separate Zr±Zr distances relative to a single 0.366 nm distance in the crystalline zircon and the loss of observable Zr±Si correlations, as mentioned previously, are both indicative of the level of distortion in the zircon structure. Clearly, the Zr sub-lattice has been signi®cantly impacted by the accumulated a-decay damage. Care needs to be taken however before interpreting the Zr K-edge EXAFS results as evidence for phase separation in the amorphous material, as has been recently proposed [26,27]. The loss of observable Zr±Si correlations on the Zr sub-lattice does not imply the di€usion of Zr and Si to form zirconia-rich and silica-rich microdomains. Indeed the signi®cant discrepancies between the Zr K-edge ®tting results from the amorphous 238 Pu-substituted zircon and monoclinic ZrO2 do not support this conclusion, and other researchers [12,23] reported no evidence for decomposition into component oxides in natural metamict zircons. Further, if signi®cant cation di€usion were to occur during the cooling of a-recoil cascades as claimed [27], then it should be equally evident for the Pu environment and not merely restricted to the Zr environment.

238

Pu-substituted zircon, along with reference data from ZrSiO4 [21] and monoclinic ZrO2

Shells

Zr±O

Zr±Zr

Zr±Zr

Distance (nm) Number D Sigma

0.216 (3) 4 (1) 0.13 (2)

0.332 (2) 4 (1) 0.13

0.408 (2) 2.8 (9) 0.13

Shells

Zr±O

Zr±O

Zr±Si

Zr±Si/Zr

Distance (nm) Number

0.213 4

0.227 4

0.299 2

0.366 4/4

Shells

Zr±O

Zr±O

Zr±Zr

Zr±Zr

Distance (nm) Number

0.210 (7) 4

0.223 (3) 3

0.346 (12) 7

0.391 2

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

The failure to observe Zr±Si correlations in amorphous zircon may be explained in terms of irregular rotation of the Zr and Si polyhedra with respect to each other, as discussed above. This would serve to destroy the unique Zr±Si distances present in crystalline zircon, and consequently prevent them from being observed by EXAFS. This is consistent with the Raman spectroscopy results [28] that indicate the SiO4 tetrahedra remain as relatively stable structural units in extremely metamict zircons but are tilted relative to each other, although there was no evidence for increased polymerization of the SiO4 tetrahedra. Such a structural rearrangement could also possibly explain the presence of two Zr±Zr distances, whose average distance is consistent with the single Zr±Zr distance in crystalline zircon. More importantly, this would preserve the zircon stoichiometry on the atomic scale, as required from the Pu LIII -edge EXAFS results. Self-radiation damage from the accumulated dose of 2:8  1019 a-decays/g in this 238 Pu-substituted zircon has clearly resulted in the loss of long-range order and formation of an amorphous structure for length scales greater than 0.5 nm. Both the PDF and EXAFS results are consistent with the interpretation of loss of longrange order, but one in which a distorted zircon structure, consisting of SiO4 and ZrO8 polyhedra rotated relative to each other, is retained over length scales of up to 0.5 nm. These observations preclude the decomposition of zircon into component oxides as a result of selfradiation damage. 3.3.

238

Pu-substituted zircon annealed at 1000°C for 12 h

Annealing the amorphous 238 Pu-substituted zircon at 1000°C for 12 h in air resulted in complete oxidation of the Pu3‡ to Pu4‡ and the crystallization of zirconÕs constituent oxides. The XRD pattern for this sample is shown in Fig. 7, along with a di€raction pattern taken from an empty Kapton holder. The multiple-layer Kapton holder required to contain these radioactive samples is responsible for the signi®cant non-linear background present in the di€raction pattern from the annealed zircon. The positions of the principal PuO2 peaks have been marked in Fig. 7, along with the main SiO2 peak; the remaining, slightly broad, di€raction peaks are from ZrO2 . The coincidence of the main SiO2 peak with the main ZrSiO4 re¯ection precludes a de®nitive statement regarding the complete absence of any zircon based on this single peak. Detailed ®tting of the broad ZrO2 di€raction maxima revealed that both the tetragonal and cubic forms are present. The lattice parameter from the cubic PuO2 is 0.5388 nm, which is considerably smaller than the 0.53960 nm reported for pure PuO2 [29], suggesting that the PuO2 may contain a proportion of ZrO2 in solid solution. As indicated in Table 1, the ionic radius of Zr4‡ is smaller than that of

219

Fig. 7. XRD pattern from the 238 Pu-substituted zircon annealed at 1000°C for 12 h in air, along with a scan from an empty Kapton holder (dotted line). The major PuO2 and SiO2 lines are indicated, all other re¯ections are from a mixture of cubic and tetragonal ZrO2 .

Pu4‡ , which would account for the lattice contraction observed for a (Pu, Zr)O2 solid solution relative to PuO2 . No evidence of di€use scattering that may be attributed to any residual amorphous material was found in the di€raction pattern, although X-ray scattering experiments are more sensitive to heavier elements and the presence of amorphous SiO2 would be dicult to detect. The PDF data obtained from this di€raction pattern are shown in Fig. 8, together with a ®t based on a mixture of cubic and tetragonal ZrO2 , PuO2 and SiO2 .

Fig. 8. PDF plot from the 238 Pu-substituted zircon annealed at 1000°C for 12 h in air, including a ®t to the data based on the crystalline component oxides of zircon. Major contributions to the total PDF from ZrO2 (cubic or tetragonal), PuO2 and SiO2 are marked by Z, P and S, respectively.

220

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

Table 5 Pu LIII -edge EXAFS ®tting results from the 238 Pu-substituted zircon annealed at 1000°C for 12 h, along with reference data for PuO2 [29] 238

Pu-zircon 1000°C/12 h

PuO2

Shells

Pu±O

Pu±O

Pu±Si

Pu±Si

Pu±Pu

Pu±Zr

Distance (nm) Number D Sigma

0.219 (2) 5 (1) 0.00 (4)

0.236 (2) 6 (1) 0.00 (4)

0.307 (2) 2.0 (6) 0.09 (2)

0.345 (2) 5 (1) 0.06 (2)

0.380 (1) 5 (1) 0.00 (3)

0.476 (1) 4 (1) 0.07 (1)

Shells

Pu±O

Pu±Pu

Distance (nm) Number

0.234 8

0.382 12

Table 6 Zr K-edge EXAFS ®tting results from the tetragonal ZrO2 [30]

238

Pu-substituted zircon annealed at 1000°C for 12 h, along with reference data for

Shells

Zr±O

Zr±O

Zr±Zr

Zr±Zr

Zr±Zr

Zr±Zr

238

Pu-zircon 1000°C/12 h

Distance (nm) Number D Sigma

0.209 (2) 3.5 (9) 0.07 (2)

0.235 (2) 1.9 (6) 0.07 (2)

0.364 (1) 9 (2) 0.06 (1)

0.516 (2) 3 (1) 0.07 (1)

0.717 (2) 14 (4) 0.09 (1)

0.872 (1) 6 (2) 0.06 (1)

ZrO2 (T)

Distance (nm) Number

0.210 4

0.236 4

0.36±0.363 12

0.51±0.518 6

0.72±0.726 12

0.887 8

While the ®t to the data is not perfect, all the major features have been accounted for. When combined with the narrow and clearly de®ned XRD features in Fig. 7, the results exclude the presence of signi®cant residual amorphous material. As expected, PuO2 dominated the Pu LIII -edge EXAFS from this 238 Pu-substituted zircon annealed at 1000°C, as shown in Table 5. Signi®cantly, the weak Pu± Si correlations observed at 0.3 and 0.35 nm are consistent with the presence of some Pu in zircon. Given the crystalline nature of the sample and the minor peak in the region of the main zircon peak, the observed Pu±Si correlations are consistent with the onset of some recrystallization of zircon. Tetragonal ZrO2 dominated the Zr K-edge EXAFS, as illustrated by the results and literature data [30] given in Table 6. Apart from the split Zr±O distances, it is very dicult to separate cubic and tetragonal ZrO2 based on the ®tting results. To summarize, annealing the 238 Pu-substituted zircon at 1000°C for 12 h in air appears to have resulted in the complete crystallization of the zirconÕs constituent oxides, as well as the presence of a small amount of Puzircon. This was quite unexpected in light of the previous isochronal step-annealing study carried out on this material at approximately half the current dose [3,5,6], in which the onset of minor ZrO2 crystallization was observed only at temperatures between 1050°C and 1200°C. In the case of natural metamict zircons, where the level of damage (i.e., accumulated dose and amount of residual crystallinity) may vary between samples, isochronal step-annealing over this same temperature

range results in the formation of ZrO2 in highly metamict zircon [11] or no ZrO2 in partially metamict zircon [15]. These results suggest that the decomposition into component oxides may depend on both the initial amorphous or metamict state (partially or fully amorphous) and the structural evolution during isochronal step-annealing. 3.4.

238

Pu-substituted zircon annealed at 1200°C for 12 h

Based on the results at 1000°C, it is not surprising that annealing the amorphous 238 Pu-substituted zircon at 1200°C for 12 h in air results not only in the complete oxidation of the Pu3‡ to Pu4‡ but also in the complete recrystallization of the zircon structure, along with some PuO2 , as shown in Fig. 9. No evidence for a residual amorphous component was present in the di€raction pattern, so the zircon had been fully recrystallized by annealing at a temperature, some 250°C lower than previously reported [3,5,6]. The lattice parameters obtained from the 238 Pu-substituted zircon annealed at 1200°C for 12 h in air are listed in Table 7. An estimate of the amount of Pu present in the zircon, based on the measured lattice parameters in the manner described above, would be between 65% and 70% of the original 0.08 formula units depending on the factor used. The recrystallized 238 Pusubstituted zircon would therefore appear to contain more Pu than the annealed crystalline 239 Pu-substituted zircon.

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

Fig. 9. XRD pattern from the 238 Pu-substituted zircon annealed at 1200°C for 12 h in air. The major PuO2 lines are indicated. All other re¯ections are from zircon.

221

Fig. 10. Comparison of the Fourier transforms of the Pu LIII edge EXAFS taken from the 238 Pu- and 239 Pu-substituted zircons after annealing in air at 1200°C for 12 h. Note that peaks in the Fourier transform have not been corrected for phase shift and as a result appear 0.05±0.10 nm shorter than ®tted values.

Table 7 A comparison of the unit cell volumes obtained from the 238 Pusubstituted zircon annealed at 1200°C for 12 h in air with those for a reference zircon [21]

Cell volume (nm3 ) D rel. ZrSiO4 Pu valence Ionic size (nm) % Pu in zircon

238

Pu-zircon 1200°C/12 h

ZrSiO4

0.2631 +0.0023 +4 0.086 (Pu) 65%

0.2608 0 ± 0.072 (Zr) 0%

A comparison of the Fourier transforms of the Pu LIII -edge EXAFS taken from the 238 Pu- and 239 Pu-substituted zircons heated at 1200°C for 12 h is shown in Fig. 10. The intensity of the transform modulus for the annealed 238 Pu-substituted zircon is greater than from the annealed 239 Pu-substituted zircon over all distances apart from where the distinguishable PuO2 distances reside. This is consistent with the 1200°C annealed 238 Pu-substituted zircon containing a greater proportion of Pu in the zircon lattice than the annealed crystalline 239 Pu-substituted zircon. The strong contribution from the annealed 238 Pu-substituted zircon at 0.5 nm is consistent with the large Pu±Zr correlations for zircon in this region. These are signi®cantly weaker in the annealed 239 Pu-substituted zircon, whose structure is dominated by the contributions from PuO2 . A comparison of the Zr K-edge Fourier transforms from the 238 Pu- and 239 Pu-substituted zircons annealed at 1200°C is shown in Fig. 11. Excellent agreement exists in both the shape and width of the transform features out to 1 nm. While intensity variations are present, the

Fig. 11. Comparison of the Fourier transforms of the Zr Kedge EXAFS taken from the 238 Pu-and 239 Pu-substituted zircons after annealing in air at 1200°C for 12 h. Note that peaks in the Fourier transform have not been corrected for phase shift and as a result appear 0.05±0.10 nm shorter than ®tted values.

equivalent shape and width of the transform features indicates that the local Zr environments in these two samples are comparable. No evidence for any residual amorphous material in the 238 Pu-substituted zircon annealed at 1200°C is evident based on the structure of the Zr sub-lattice. The equivalency of the Zr environments observed in the 238 Pu-substituted zircon annealed at 1200°C with those in both 239 Pu-substituted zircons annealed at 1200°C (Fig. 11) and reference zircon [21] (Table 8) is consistent with the recrystallization of the amorphous

222

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

Table 8 Zr K-edge EXAFS ®tting results from the ZrSiO4 [21] 238

Pu-zircon 1200°C/12 h

ZrSiO4

238

Pu-substituted zircon annealed at 1200°C for 12 h in air, along with reference data for

Shells

Zr±O

Zr±Si

Zr±Si/Zr

Zr±Si

Zr±Zr

Distance (nm) Number D Sigma

0.216 (2) 8 (2) 0.10 (2)

0.301 (2) 1.6 (5) 0.05 (2)

0.364 (1) 1.6 (5) 0.0

0.469 (2) 3 (1) 0.0

0.557 (1) 6 (2) 0.04 (2)

Shells

Zr±O

Zr±O

Zr±Si

Zr±Si/Zr

Zr±Si

Zr±Zr

Distance (nm) Number

0.213 4

0227 4

0.299 2

0.366 4/4

0.467 4

0.557 5

238

Pu-substituted zircon (at 1200°C) directly from the amorphous state and not via the formation of oxide intermediaries. Further insight is provided by the nature of the PuO2 that forms during annealing at 1200°C. The PuO2 exsolved from the crystalline 239 Pu-substituted zircon at 1200°C is expected to be pure and free of any ZrO2 . If the amorphous 238 Pu-substituted zircon annealed at 1200°C recrystallizes via the formation of oxide-intermediaries, the PuO2 present in the annealed sample is expected to contain some ZrO2 , similar to that observed in the 238 Pu-substituted zircon annealed at 1000°C. However, if zircon recrystallizes directly from the amorphous state, the PuO2 that forms is expected to be pure. The cubic PuO2 lattice parameters from both the 238 Pu- and 239 Pu-substituted zircons are equivalent with pure PuO2 [29], being 0.5400, 0.5393 and 0.5396 nm, respectively. This indicates that no ZrO2 is present in the PuO2 , which suggests that zircon recrystallizes directly from the amorphous state at 1200°C. This deduction is also consistent with the previous annealing study of this material [3,5,6], where annealing at 1450° for 12 h was required to recrystallize the zircon structure because phase separation occurred at lower temperatures during the step-annealing process. A comparison of the PDF data obtained from the XRD patterns taken from the 238 Pu- and 239 Pu-substituted zircons after their respective anneals at 1200°C for 12 h is shown in Fig. 12. Excellent structural agreement is evident between these two samples. The intensities of the Zr±Zr/Pu pairs in the annealed 238 Pu-substituted zircon, as indicated in Fig. 12, are generally more intense than in the annealed 239 Pu-substituted zircon, which is consistent with the incorporation of a greater proportion of Pu in the recrystallized 238 Pu-substituted zircon. No evidence for any residual amorphous material in the 238 Pu-substituted zircon annealed at 1200°C for 12 h in air is evident from the PDF data. The ability on one hand to crystallize the constituent oxides of zircon at 1000°C, and on the other to fully recrystallize zircon at 1200°C, some 250°C lower than previously reported, is very unexpected and raises important questions as to how the current ®ndings relate to the previous annealing studies discussed earlier. The

Fig. 12. A comparison of the PDF data from the 238 Pu- and 239 Pu-substituted zircons annealed at 1200°C for 12 h in air. The positions of the principal Zr±Zr/Pu pairs in the Pu-substituted zircons are indicated by arrows.

principal di€erence between the current study and previous studies is the manner in which the samples were annealed. In previous isochronal annealing studies of highly amorphous zircons [3,5,6,11], a single sample was sequentially heated to successively higher temperatures; whereas in the current study, separate samples were rapidly ramped (600°C/h) to the designated annealing temperature. In the previous study [3,5,6] of the amorphous 238 Pu-subsubstituted zircon, a two-stage recrystallization process, via the formation of oxide intermediaries between 1050°C and 1200°C, was observed, with complete recovery of the crystalline zircon structure at 1450°C. In the current study of the same material at a signi®cantly higher dose, the amorphous state of the 238 Pu-substituted zircon is directly and fully recrystallized to the zircon structure at temperatures as low as 1200°C, but only because it is ramped rapidly to the annealing temperature. This suggests that rapid heating through the temperature regime where decomposition of the amorphous state to component oxides is observed can lead to congruent recrystallization of the zircon structure. However, if intermediate oxides are

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

allowed to form, then much higher annealing temperatures are required for recovery of the zircon crystal structure. It should be noted that in an isochronal annealing study of partially metamict zircons [15], which contained both amorphous and damaged-crystalline regimes, no intermediate formation of ZrO2 was observed during recrystallization of the amorphous domains, and recovery of the crystalline zircon structure was complete at 900°C. Thus, structure evolution during annealing is not only dependent on the initial structural state and subsequent annealing temperature, but also on the heating rate and time at temperature. It is the interplay of the kinetics of various recovery and di€usion processes that control structural evolution during annealing. One of the keys to understanding this behavior lies in characterizing the short-range structure of the amorphous state. The general inference from previous annealing studies [3,5,6,11], as illustrated in Fig. 1, is that during the isochronal step-anneals up to 900°C, where no signi®cant recovery of density or recrystallization was observed, no signi®cant change in the amorphous structure had occurred. However, the current results suggest that this is not the case. While no long-range order has developed, signi®cant alteration to the shortrange structure has probably occurred that impacts subsequent recrystallization behavior. Closer examination of the previous density recovery data (Fig. 1) indicates that some decrease in density occurs between 600°C and 900°C. This temperature range coincides with that where both increases in the mean Zr±O and Zr±Zr distances and an increase in Zr coordination number are observed in the EXAFS data for annealed metamict zircons [15]. Furthermore, perturbed angular correlation spectroscopy (PACS) data [17] show that local ordering occurs in metamict natural zircons for annealing above 600°C. Thus, the initial amorphous structure of zircon, which is devoid of long-range order but retains a distorted zircon structure and stoichiometry over length scales up to 0.5 nm, begins short-range structural evolution at temperatures above 600°C. This onset of shortrange structural changes precedes the decomposition into component oxides observed in the isochronal stepannealing studies [3,5,6,11] and may precede the decomposition observed for the rapid-ramp annealing at 1000°C in the current study. The key to the direct recrystallization of amorphous zircon at 1200°C is rapidly heating the sample through the temperature range where decomposition is apparently energetically favored. The question that remains unanswered is ``why is decomposition into component oxides favored over recrystallization of the zircon structure at low temperatures?'' Both step-annealing and rapid-ramp annealing at temperatures below 1200°C appear to result in the decomposition of amorphous zircon into component oxides (crystalline ZrO2 and amorphous SiO2 ). This implies that the distorted zircon structure and stoichi-

223

ometry that is evident in the amorphous structure at room temperature is lost during structural evolution at intermediate temperatures. The phase separation associated with the observed decomposition implies sucient cation di€usion for nucleation and growth of ZrO2 , which implies that the decomposition will be kinetically limited. Thus, as observed, rapid heating to higher temperatures can negate such decomposition. However, this does not answer the question as to why the amorphous structure will preferentially decompose rather than recrystallize, particularly since the local composition and structure is that of distorted zircon. Both the decomposed constituent oxide structure and the recrystallized zircon structure represent lower energy states than the amorphous state [16], with crystalline zircon being the lowest energy state in the temperature range of this study. Consequently, there is an energetic driving force for both structural rearrangements; however, there is apparently a lower energy barrier for decomposition (i.e., cation di€usion) than for recrystallization of the zircon structure. Since the SiO4 tetrahedra geometry is stable in metamict zircon [28] and is almost temperature independent in zircon [31], most short-range structural changes with increasing temperature are associated with observed increases in both Zr± O distances and Zr coordination, which may a€ect or enhance Zr di€usion. Although tetravalent cation selfdi€usion in crystalline zircon has exceedingly high activation energies (8 eV) [32,33], this is largely due to the very high formation energy for the Zr vacancy, as discussed in a recent computer simulation study [34] that determined the migration energy for the Zr vacancy to be 1.2±1.4 eV in crystalline zircon. Consequently, in highly damaged or amorphous zircon, where there is an existing high concentration of Zr vacancies or where vacancy formation energies may be considerably reduced, self-di€usion may be activated at lower temperatures. One might further speculate that the increase in Zr±O distances with temperature may increase the mobility of Zr in the amorphous zircon structure, particularly while the coordination number is still less than eight. This provides some explanation for the observed preferential decomposition of amorphous zircon to constituent oxides via Zr cation di€usion at intermediate temperatures. The implication is that the energy barrier for recrystallization of zircon is higher than for Zr selfdi€usion in the amorphous state. Some insight into this is provided by the work of Mursic and coworkers [14], who observed that recrystallization of the zircon structure occurs in a temperature range where the sudden expansion of the Si±O bond and SiO4 tetrahedral volume allows a Zr±O polyhedral subunit to rotate causing a structural relaxation. From the results discussed in this paper, the energy barrier for rotation of the Zr±O polyhedra must obviously be larger than the energy barrier for Zr self-di€usion in the amorphous state.

224

B.D. Begg et al. / Journal of Nuclear Materials 278 (2000) 212±224

Future computer simulations may provide a means to con®rm these observations. 4. Conclusions The oxidation of Pu3‡ to Pu4‡ in the crystalline Pu-substituted zircon results in a signi®cant decrease in lattice distortion due to the reduction in ionic size mismatch between the Pu3‡ and Pu4‡ on the Zr4‡ site. The nature of the ®rst Zr±Si correlation in zircon is very sensitive to the level of lattice distortion present in the zircon. In addition, a signi®cant amount of PuO2 is exsolved from the zircon lattice after heating at 1200°C in air. Detailed characterization of the amorphous 238 Pusubstituted zircon is consistent with the loss of both long-range order and edge-sharing relationships between SiO4 and ZrO8 polyhedra. However, the amorphous state retains a distorted zircon structure and stoichiometry, consisting of SiO4 and ZrO8 polyhedra rotated relative to each other, over length scales of up to 0.5 nm. The recrystallization of the amorphous 238 Pu-substituted zircon could be achieved directly at temperatures as low as 1200°C if heated rapidly through the intermediate temperature regime where decomposition to oxides is preferred. The decomposition of amorphous zircon to constituent oxides is probably kinetically limited by Zr di€usion, which has a lower energy barrier than the polyhedral rotation required for recrystallization of the zircon structure from the amorphous state.

239

Acknowledgements This study was supported by the Environmental Management Science Program, Oce of Environmental Management, US Department of Energy under Contract DE-AC06-76RLO 1830. The XAS and XRD experiments were conducted at the Stanford Synchrotron Radiation Laboratory, which is supported by Oce of Basic Energy Sciences, US Department of Energy. The assistance of the Seaborg Institute for Transactinium Science at Los Alamos National Laboratory in providing Health Physics and other experimental support is gratefully acknowledged. References [1] R.C. Ewing, W.J. Weber, F.W. Clinard Jr., Progr. Nucl. Energy 29 (2) (1995) 63. [2] W.J. Weber, R.C. Ewing, C.R.A. Catlow, T. Diaz de la Rubia, L.W. Hobbs, C. Kinoshita, Hj. Matzke, A.T. Motta, M. Natasi, E.K.H. Salje, E.R. Vance, S.J. Zinkle, J. Mater. Res. 13 (1998) 1434.

[3] W.J. Weber, J. Mater. Res. 5 (1990) 147. [4] T. Murakama, B.C. Chakoumakos, R.C. Ewing, G.R. Lumpkin, W.J. Weber, J. Am. Mineral. 76 (1991) 1510. [5] W.J. Weber, J. Am. Ceram. Soc. 76 (1993) 1729. [6] W.J. Weber, R.C. Ewing, L.M. Wang, J. Mater. Res. 9 (1994) 688. [7] N.J. Hess, W.J. Weber, S.D. Conradson, J. Nucl. Mater. 254 (1998) 175. [8] N.J. Hess, W.J. Weber, S.D. Conradson, J. Alloys Comp. 271±273 (1998) 240. [9] R.D. Shannon, Acta Crystallogr. B 25 (1969) 925. [10] R.E. Williford, B.D. Begg, W.J. Weber, N.J. Hess, this issue, p. 207. [11] E.R. Vance, B.W. Anderson, Mineral. Mag. 38 (1972) 605. [12] R.L. Barinsky, I.M. Kulikova, Phys. Chem. Miner. 1 (1977) 325. [13] I. Nakai, J. Akimoto, M. Imafuku, R. Miyawaki, Y. Sugitani, K. Koto, Phys. Chem. Miner. 15 (1987) 113. [14] Z. Mursic, T. Vogt, H. Boysen, F. Frey, J. Appl. Crystallogr. 25 (1992) 519. [15] F. Farges, Phys. Chem. Miner. 20 (1994) 504. [16] S. Ellsworth, A. Navrotsky, R.C. Ewing, Phys. Chem. Minerals 21 (1994) 140. [17] H. Jaeger, L. Abu-Radad, D.J. Wick, Appl. Radiat. Isot. 48 (1997) 1083. [18] E.R. Vance, B.W. Anderson, Mineral. Mag. 38 (1972) 721. [19] D.A. Dimitrov, A.L. Ankudinov, A.R. Bishop, S.D. Conradson, Phys. Rev. B 58 (1998) 14227. [20] H.P. Klug, L.E. Alexander, X-ray Di€raction Procedures for Polycrystalline and Amorphous Materials, Wiley, New York, 1974, p. 791. [21] Powder Di€raction File PDF-2 Database, PDF 6-266, JCPDS-ICDD, 1998. [22] C. Keller, Nukleonik 5 (1963) 41. [23] F. Farges, G. Calas, Am. Mineral. 76 (1991) 60. [24] F. Farges, R.C. Ewing, G.E. Brown Jr., J. Mater. Res. 8 (1993) 1983. [25] R.E. Hann, P.R. Suitch, J.L. Pentecost, J. Am. Ceram. Soc. 68 (1985) 285. [26] A.C. McLaren, J.D. Fitz Gerald, I.S. Williams, Geochim. Cosmochim. Acta 58 (1993) 993. [27] A. Meldrum, S.J. Zinkle, L.A. Boatner, R.C. Ewing, Nature 395 (1998) 56. [28] L. Nasdala, G. Irmer, D. Wolf, Eur. J. Mineral 7 (1995) 471. [29] Powder Di€raction File PDF-2 Database, PDF 41-1170, JCPDS-ICDD, 1998. [30] J.C. Howard, E.H. Kisi, R.B. Roberts, R.J. Hill, J. Am. Ceram. Soc. 73 (1990) 2828. [31] R.M Hazen, L.W. Finger, Temperature, Pressure, Composition and the Variation of Crystal Structure, Wiley, New York, 1982, p. 231. [32] D.J. Cherniak, J.M. Hanchar, E.B. Watson, Contrib. Minerol. Petrol. 127 (1997) 383. [33] D.J. Cherniak, J.M. Hanchar, E.B. Watson, Chem. Geol. 134 (1997) 289. [34] R.E. Williford, W.J. Weber, R. Devanathan, A.N. Cormack, J. Nucl. Mater. 273 (1999) 164.